1. Field of the Invention
The present invention relates to electronic exchanges with a conference telephone communication function and conference communication methods which add telephone communication signals on speakers in time sharing multiplexed channels, subtract the telephone communication signal on any particular speaker and transfer them to the remaining speakers, and more particularly to apparatus and methods which prevent the deterioration of the voice quality occurring, for example, when a plurality of speakers start to talk simultaneously.
2. Description of the Related Art
Generally, methods of exchanging a conference telephone communication (simultaneous telephone communications by three or more speakers) through an electronic exchange are mainly divided into two: an "adding method" which adds all the speaker' voice signals and transfers the result to the respective speakers and an "exchange method" which selects a particular speaker's voice signal and transfers it to the respective speakers.
The "adding method" will be now described. The "adding method" is a so-called (N-1) adding method which is a system in which the telephone communication signals (PCM data) on the voices of all the speakers participating in a conference are temporarily added, the telephone communication signal (PCM data) of a particular speaker is subtracted from the result of the addition and the result of the subtraction is transferred to the respective speakers. The data (PCM data) thus formed is transferred as telephone communication signals to the respective speakers (participating in the conference).
When telephone communication signals including the PCM data are transmitted to the respective participators, they are converted via a digital-to-analog (D/A) converter to analog signals, which are then demodulated as voices.
In such a conference telephone communication through an electronic exchange, two or more speakers may often talk simultaneously because the conference is a simultaneous telephone communication by three or more speakers.
Therefore, the signal levels of the acoustically demodulated analog signals are unstable. Even if a voice signal at a preferred level is obtained when one speakers talks, the signal level obtained if two or more speakers start to talk simultaneously would automatically become high enough to be offensive to the ear. In an extreme case, the signal becomes distorted to thereby greatly deteriorate the voice quality.